A Theoretical Development and Analysis of Jumping Gene Genetic Algorithm.
ABSTRACT Recently, gene transpositions have gained their power andattentionsincomputationalevolutionaryalgorithmdesigns.In 2004, the Jumping Gene Genetic Algorithm (JGGA) was first pro- posedandtwonewgenetranspositionoperations,namely,cut-and- paste and copy-and-paste, were introduced. Although the outper- formance of JGGA has been demonstrated by some detailed statis- tical analyses based on numerical simulations, more rigorous the- oretical justification is still in vain. In this paper, a mathematical model based on schema is derived. It then provides theoretical jus- tifications on why JGGA is superiority in searching, particularly when it is applied to solve multiobjective optimization problems. The studies are also further verified by solving some optimization problems and comparisons are made between different optimiza- tion algorithms.
- SourceAvailable from: Ke Li
[Show abstract] [Hide abstract]
- "Instead of using the JG operators directly, RJGGA just employs the polynomial mutation  and Simulated Binary Crossover (SBX)  to emulate the single and double chromosomes' JG operators, respectively. Recently, Tang et al.  theoretically justified the outperformance of binary-coded JG. Besides the theoretical studies, JG has also been widely applied to various real-world optimization scenarios, such as passive circuits systems , planar UWB monopole antenna design , power voltage control systems  and wireless local area network optimization . "
ABSTRACT: Exploration and exploitation are two cornerstones of evolutionary multiobjective optimization. Most of the existing works pay more attention to the exploitation, which mainly focuses on the fitness assignment and environmental selection. However, the exploration, usually realized by traditional genetic search operators, such as crossover and mutation, has not been fully addressed yet. In this paper, we propose a general learning paradigm based on Jumping Genes (JG) to enhance the exploration ability of multiobjective evolutionary algorithms. This paradigm adapts the JG to the continuous search space, and its activation is completely adaptive during the evolutionary process. Moreover, in order to efficiently utilize the useful information, only non-dominated solutions eliminated by the environmental selection are chosen for the secondary exploitation. Empirical studies demonstrate that the performance of a baseline algorithm can be significantly improved by the proposed paradigm.Information Sciences 03/2013; 226:1-22. DOI:10.1016/j.ins.2012.11.002 · 3.89 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Layout planning in a manufacturing company is an important economical consideration. In the past, research examining the facility layout problem (FLP) generally concerned static cases, where the material flows between facilities in the layout have been assumed to be invariant over time. However, in today’s real-world scenario, manufacturing system must operate in a dynamic and market-driven environment in which production rates and product mixes are continuously adapting. The dynamic facility layout problem (DFLP) addresses situations in which the flow among various facilities changes over time. Recently, there is an increasing trend towards implementation of industrial robot as a material handling device among the facilities. Reducing the robot energy usage for transporting materials among the facilities of an optimal layout for completing a product will result in an increased life for the robots and thus enhance the productivity of the manufacturing system. In this paper, we present a hybrid genetic algorithm incorporating jumping genes operations and a modified backward pass pair-wise exchange heuristic to determine its effectiveness in optimizing material handling cost while solving the DFLP. A computational study is performed with several existing heuristic algorithms. The experimental results show that the proposed algorithm is effective in dealing with the DFLP.Paladyn 01/2011; 2(3):164–174. DOI:10.2478/s13230-012-0008-1